Most spectrum analyzer systems for analyzing microwave and RF spectra function by mixing a signal of interest (the signal to be analyzed) with an RF carrier, creating a lower frequency signal with power proportional to the incoming RF signal. The RF carrier is swept in time such that this lower frequency signal is swept over the narrow passband of an electronic filter. The power passing through the filter is measured as the carrier is swept in order to reconstruct the spectral power of the signal of interest. The disadvantage of this system is that only a small portion of the signal of interest can be observed instantaneously. In order to overcome this difficulty, many attempts have been made to create an optical spectrum analyzer in which a wideband electronic signal is converted to light. The light pattern representing the signal is then viewed in its entirety.
Two of the present inventors, along with two others, invented a microwave camera which was patented under U.S. Pat. No. 5,121,124 issued Jun. 9, 1992. In one of the described embodiments of that invention, a signal of interest is sent to a Bragg Cell (an acousto-optic modulator) which, when illuminated with a laser, diffracts a portion of the laser beam onto the pixel array of a video camera. The image of the diffracted light represents the spectrum of the signal of interest. This use of acousto-optic modulators (AOMs) as optical spectrum analyzers has been quite successful in providing high resolution and real-time operation (see also U.S. Pat. No. 4,633,170 to Burns). Unfortunately, AOMs are currently limited to processing a bandwidth of about 2 GHz.
Electro-optic modulators, however, are currently available with bandwidths of over 20 GHz. Their use in spectral analyzer systems would allow for the instantaneous analysis of signals with much broader bandwidths. One approach has been to modulate a light beam such as a laser beam with an RF or microwave signal then to use dispersive elements such as diffraction gratings (U.S. Pat. No. 4,464,624 to Osterwalder) to spatially separate the different frequency components in the modulated beam. The drawback of such dispersive element systems is that the angular diffraction between different frequency components is extremely small. To obtain adequate resolution, the size of such systems must be on the order of several meters.
Another approach (U.S. Pat. No. 4,871,232 to Grinberg, et. al.) involves a time sampling of the modulated optical beam rather than a spatial sampling. Using two parallel plates as an optical waveguide, a collimated beam is sampled at regularly spaced intervals by making one of the plates totally reflective and the other partially reflective. The various beam samples "steer" the beam, much like a frequency scanned antenna. This approach becomes impractical in that the beam at each sampling point must possess a small dimension and approximate intensity of the first sampled point if the beam steering is to be effective. Similar approaches, such as a group of fiber optic delay lines with varying delays, exhibit similar difficulties in the precise tolerances involved.
Another proposed spectrum analyzer system uses an electro-optic modulator in conjunction with a swept Fabry-Perot etalon (U.S. Pat. No. 5,041,778). The etalon acts as a very narrow band optical filter. Here, the spacing of the etalon is swept in time, varying its resonance characteristics and the frequency that is allowed to pass through the etalon. Only one detector is used to measure transmitted power. This device suffers from the same limitations as electronic spectrum analyzers in that the entire spectrum of interest cannot be viewed simultaneously.